Evaluation of commercial latex reagents for identification of O157 and H7 antigens of Escherichia coli.

Agglutination reactions obtained with three commercial latex reagents for detecting Escherichia coli O157 antigen (Oxoid Diagnostic Reagents, Hampshire, England; Pro-Lab Inc., Richmond Hill, Ontario, Canada; and Remel Microbiology Products, Lenexa, Kans.) and one for detecting H7 antigen (Remel) were compared with those obtained by standard serologic methods by using the Centers for Disease Control and Prevention (CDC) reference antisera for O157 and H7 antigens. For 159 strains of E. coli and related organisms, the Oxoid, Pro-Lab, and Remel O157 latex reagents each had a sensitivity and specificity of 100% compared with the CDC reference antiserum. For 106 strains of E. coli and related organisms that were not enhanced for motility through semisolid medium, the Remel H7 latex reagent had a sensitivity of 96% and a specificity of 100% compared with the standard tube agglutination method with CDC H7 antiserum. Measures to enhance motility were needed for some strains to detect the H7 antigen. Our findings demonstrate that the commercial latex reagents are good alternatives to standard serologic methods for identifying the O157 and H7 antigens of E. coli.

Isolation and characterization of a beta-D-glucuronidase-producing strain of Escherichia coli serotype O157:H7 in the United States.

A phenotypic variant of Escherichia coli serotype O157:H7 (G5101) was isolated from a patient with bloody diarrhea. Strain G5101 does not ferment sorbitol but is beta-D-glucuronidase and urease positive. Serotyping and colony hybridization using a serotype-specific DNA probe confirmed that the isolate was O157:H7. G5101 produces Shiga-like toxins I and II and contains an eae gene that is highly conserved in the O157:H7 serotype. This strain would have been missed by laboratories that screen for the sorbitol-negative, beta-D-glucuronidase-negative phenotype in isolating E. coli O157:H7 from clinical and food specimens.

Molecular subtyping of Salmonella enteritidis phage type 8 strains from the United States.

Salmonella enteritidis is now the most common serotype of the genus Salmonella reported in the United States. Bacteriophage typing has been helpful for subdividing S. enteritidis strains from different sources in the United States. Most S. enteritidis outbreaks reported were egg related, and the majority of them were caused by strains of phage type 8. To determine whether restriction fragment length polymorphism of the rRNA genes (ribotyping) and of the genomic DNAs from two lysogenic phages from S. enteritidis could be used to discriminate between S. enteritidis phage type 8 strains, we conducted Southern hybridization studies on 24 isolates from different outbreaks and six non-outbreak-associated strains using DNA probes for 16S and 23S rRNA genes and S. enteritidis typing phages 1 and 2 from the Ward typing system (L. R. Ward, J. D. H. de Sa, and B. Rowe, Epidemiol. Infect. 99:291-294, 1987). Of seven restriction endonucleases screened with the probe for rRNA genes, AccI provided the best discrimination between strains; six distinct patterns were observed. AccI ribosomal DNA patterns 1 to 6 were detected among 76.7, 3.3, 6.7, 3.3, 3.3, and 6.7% of isolates tested, respectively. Strains of AccI ribosomal DNA pattern 3 could be further subdivided into two additional patterns by using SmaI. Epidemiologically related strains had identical patterns. No discrimination between strains was achieved by probes for phages 1 and 2. No sequences homologous to the phage I probe were detected among phage type 8 strains, and all strains tested with six restriction enzymes had the same hybridization pattern with the phage 2 probe. These findings demonstrate that ribotyping with AccI and SmaI provides an additional means of discriminating between some phage type 8 strains; however, ribotyping and the phage 2 hybridization results from egg-related outbreak strains support previous findings that these strains are closely related.

Evaluation of a technique for identification of Shiga-like toxin-producing Escherichia coli by using polymerase chain reaction and digoxigenin-labeled probes.

A polymerase chain reaction (PCR) technique for the identification of Shiga-like toxin (SLT)-producing Escherichia coli was assessed by using 95 strains of SLT-producing E. coli and 5 Shigella dysenteriae type 1 strains. PCR was used for the amplification of slt gene sequences from whole bacterial colonies. A digoxigenin-labeled DNA probe was used for identification of the PCR products in a spot blot hybridization assay. Modifications were made to adapt this technique for the proper identification of 10 SLT-producing isolates which were refractory to the heat lysis step that was used to liberate whole-cell DNA for PCR and 6 isolates which gave nonspecific amplification products. The sensitivity and specificity of this assay were each 99% when compared with toxin neutralization results by using SLT-specific monoclonal antibodies. These values indicate that this detection technique could be suitable for use in a clinical laboratory.

Isolation of a Citrobacter freundii strain which carries the Escherichia coli O157 antigen.

A biochemically typical strain of Citrobacter freundii which carries the Escherichia coli O157 antigen is described. The significance of differentiating such strains from typical E. coli O157 strains is stressed.

Serological cross-reactions between Escherichia coli O157 and other species of the genus Escherichia.

The antigenic relatedness of Escherichia coli O157 and four sorbitol-negative species of the genus Escherichia was examined. Isolates of Escherichia hermannii, E. fergusonii, E. vulneris, and E. blattae were tested in the tube agglutination assay by using polyclonal antisera and in the slide agglutination assay by using latex reagents. Only four isolates (17%) of E. hermannii exhibited serological cross-reactivity.

DNA probe analysis of diarrhoeagenic Escherichia coli: detection of EAF-positive isolates of traditional enteropathogenic E. coli serotypes among Bangladeshi paediatric diarrhoea patients.

Escherichia coli isolates from all surveillance patients less than or equal to 20 months of age seen for diarrhoea at the Dhaka Clinical Treatment Facility of the International Centre for Diarrhoeal Disease Research, Bangladesh between March 1 and August 31, 1988, were collected and hybridized with DNA probes to assess the potential importance of diarrhoeagenic E. coli among paediatric patients in Bangladesh. Of 396 patients evaluated, 18% were infected with enteropathogenic E. coli (EPEC) adherence factor (EAF)-positive E. coli, 23% were infected with enterotoxigenic E. coli (ETEC), 9% were infected with Shiga-like toxin-positive E. coli, and 13% were infected with diffuse adhesiveness-positive E. coli. None were infected with enteroinvasive E. coli. Ten percent of patients were colonized with more than one type of potential diarrhoeagenic E. coli. The majority of EAF-positive isolates were of traditional EPEC O:H serotypes. Although this was not a case-control study, the large number of EPEC and ETEC, which are recognized enteric pathogens, suggests these organisms are important causes of diarrhoeal diseases in this pediatric population.

Restriction fragment length polymorphisms in rRNA operons for subtyping Shigella sonnei.

Shigella sonnei is the most frequent cause of shigellosis in the United States. Epidemiologic studies of this organism have been hampered by the lack of adequate typing procedures. Ribosomal DNA analysis (ribotyping), a method which analyzes restriction fragment length polymorphisms in the chromosomal genes that encode rRNA, has recently been shown to be useful for microbial species identification and subtyping. To determine whether ribotyping could be used to distinguish between S. sonnei isolates, we conducted Southern hybridization studies on isolates from 16 different geographic locations and from four recent outbreaks. S. sonnei genomic DNA fragments generated following digestion with SalI hybridized with Escherichia coli 16S and 23S rRNAs to produce six distinct patterns; strains with patterns 1, 2, and 3 were each further subdivided into two additional patterns by using PvuII, SmaI, and SstI, respectively. Epidemiologically related strains had identical patterns. Ribotyping appears to be a useful tool for epidemiologic studies of shigellosis caused by S. sonnei.

Molecular epidemiologic techniques in analysis of epidemic and endemic Shigella dysenteriae type 1 strains.

During 1988 the number of Shigella dysenteriae type 1 infections reported in the United States increased fivefold. To determine if recent isolates from Mexico were related to those that caused epidemics of dysentery worldwide, Southern hybridization analysis was done with Shiga toxin and ribosomal RNA gene probes. Western hemisphere and Eastern Hemisphere strains differed by the size of a single EcoRI fragment carrying the Shiga toxin genes. Three ribosomal DNA (rDNA) patterns were observed, which correlated with the strain's continental origin for 81 of 83 isolates tested. Together the Shiga toxin and rDNA probe results indicated that recent Mexican isolates were chromosomally similar to earlier Central American isolates and distinct from Asian and African strains. This suggests there has been no significant exchange of organisms between continents in recent decades and that the 1988 outbreak in Mexico was caused by strains present in Central America since at least 1962.

The use of plasmid profiles and nucleic acid probes in epidemiologic investigations of foodborne, diarrheal diseases.

The application of nucleic acid analyses to investigations of infectious disease outbreaks has resulted in useful molecular strain markers that distinguish the epidemic clone of a particular pathogen and help identify specific vehicles of infection. We have successfully used plasmid profile analysis, restriction endonuclease digestion of plasmid and whole-cell DNAs, and nucleic acid hybridization to investigate recent outbreaks of foodborne diarrheal illness. Plasmid analysis has been important in identifying epidemic strains of Salmonella enteritidis and Escherichia coli O157:H7. In a culture survey of S. enteritidis isolates from humans and a variety of animals, including chickens and chicken eggs, we identified 16 distinct plasmid profiles and used these to differentiate strains, especially within commonly occurring phage types (Colindale 8 and 13a). HindIII digests of plasmid DNA were useful in distinguishing plasmids of similar mass but dissimilar enzyme target sequences; they clearly distinguished S. enteritidis strains causing systemic infections in children in parts of Africa from U.S. isolates. Investigations of outbreaks of hemorrhagic colitis have also been assisted by plasmid analysis. Restriction endonuclease digests of whole-cell DNA and Southern blot analysis, hybridizing with E. coli 16S and 23S rRNA (ribotyping), have been effective subtyping techniques, especially for plasmidless isolates of Campylobacter jejuni. In five outbreaks of C. jejuni infections, ribotyping of PvuII and ClaI digests distinguished individual epidemic strains within one commonly occurring C. jejuni serotype (Penner 2, Lior 4). Preliminary data show that ribotyping of NcoI digests can also distinguish individual epidemic strains of E. coli O157:H7 and may provide a more stable marker than plasmid profiles. Specific DNA probes derived from cloned virulence genes of E. coli have been invaluable in epidemic investigations and surveys. Using colony hybridization, we found in one survey of stool specimens from 174 dairy cattle that 11% of animals were asymptomatically carrying Shiga-like toxigenic E. coli other than O157:H7. We also found that newly synthesized oligonucleotide probes for the Shiga-like toxins I and II agreed 100% with cloned gene probes in a study of 613 E. coli strains. Future studies of these organisms will include the use of additional synthetic oligonucleotides as primers to amplify the toxin genes directly in patient and animal specimens by the polymerase chain reaction. There is a continuing and expanding role for molecular approaches in epidemiological investigations. The DNA methods described above are not based on the often complex expression of phenotypic characteristics, and, unlike sensitive and specific techniques such as phage typing, a single method can be used to study a variety of Gram-positive and negative bacterial pathogens.

Characterization of Shigella dysenteriae serotypes 11, 12, and 13.

We conducted serologic and biochemical studies on strains of three provisional Shigella serotypes. Included were 19 strains of serotype 3873-50 from three countries (1 from Mexico, 1 from Netherlands Antilles, and 17 from the United States), 13 strains of serotype 3341-55 from three countries (1 from Canada, 1 from Bulgaria, and 11 from the United States), and 19 strains of serotype 19809-73 from two countries (16 from Israel and 3 from the United States). Reactions of these strains with homologous and heterologous Shigella antisera showed that serotypes 3873-50 and 19809-73 were unique. Strains of serotype 3341-55 shared an identical antigen with Shigella boydii 18; however, they were distinguished from S. boydii 18 biochemically and with absorbed sera. Strains of the three serotypes were biochemically most similar to Shigella dysenteriae. Like other shigellae, strains from each serotype were invasive in the Serény assay. Although biochemically similar to S. dysenteriae, none of the provisional serotypes produced Shiga toxin or the biologically related toxin, Shiga-like toxin II, produced by some strains of Escherichia coli. On the basis of these results and other published findings, we support earlier proposals to add the three provisional serotypes 3873-50, 3341-55, and 19809-73 to the Shigella schema and will include them as S. dysenteriae 11, 12, and 13, respectively.

Evidence for participation of the macrophage in Shiga-like toxin II-induced lethality in mice.

Seven strains of inbred mice were compared for their susceptibility to the lethal effects of Shiga-like toxin II (SLT II). A/J mice, which are unable to produce the C5 component of complement, did not differ from C5 normal mice in susceptibility to SLT II. CBA/NJ mice (hemizygous for X-linked immunodeficiency) did not differ from the B-cell sufficient CBA/J strain. C3H/HeJ mice, defective in macrophage response to lipopolysaccharide (Lpsd), showed a consistently and significantly longer mean time to death than did the normally responsive C3H/HeN strain. C57BL/10ScN mice, which also carry the Lpsd allele, showed a similar but smaller difference in mean time to death compared with the C57BL/10SnJ strain. Production of tumor necrosis factor could be induced in vitro by SLT II treatment of C3H/HeN, but not C3H/HeJ macrophages. These results imply that antibody and complement production do not modulate SLT II lethality in mice, but that the macrophage may contribute to SLT II-induced injury.

Affinity purification and characterization of Shiga-like toxin II and production of toxin-specific monoclonal antibodies.

Shiga-like toxin (SLT-II) was purified to apparent homogeneity from Escherichia coli K-12 strain NM522 containing the cloned toxin genes on recombinant plasmid pEB1. Purification was accomplished by a series of column chromatography techniques: anion-exchange, chromatofocusing, cation-exchange, and monoclonal antibody affinity chromatography. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of the pure toxin showed that SLT-II consisted of A and B subunits with apparent molecular weights of 32,000 and 10,200 +/- 800, respectively. A band of molecular weight 25,000 was also observed after sodium dodecyl sulfate-polyacrylamide gel electrophoresis and identified as the A1 subunit by Western immunoblot analysis with toxin-specific monoclonal antibodies (MAbs). The pI of the purified toxin was 5.2. Approximately 1 pg of pure SLT-II, but was not neutralized by polyclonal antibodies or MAbs to SLT-I. Five hybridomas against SLT-II were produced (BC5 BB12, DC1 EH5, EA5 BA3, ED5 DF3, and GB6 BA4). Culture supernatant fluids containing MAbs from these hybridomas did not neutralize the cytotoxicity of SLT-I or Shiga toxin. Western blot analysis showed that two MAbs (MAb DC1 EH5 and MAb GB6 BA4) recognized the A and A1 subunits of SLT-II and three MAbs (MAb BC5 BB12, MAb EA5 BA3, and MAb ED5 DF3) recognized the B subunit of SLT-II. MAb BC5 BB12 was used to prepare an affinity column for toxin purification.

Cloning and sequencing of the genes for Shiga toxin from Shigella dysenteriae type 1.

The structural genes for Shiga toxin, designated stx A and stx B, were cloned from Shigella dysenteriae type 1 3818T, and a nucleotide sequence analysis was performed. Both stx A and stx B were present on a single transcriptional unit, with stx A preceding stx B. The molecular weight calculated for the processed A subunit was 32,225, while the molecular weight of the processed B subunit was 7,691. Comparison of the nucleotide sequences for Shiga toxin and Shiga-like toxin I (SLT-I) from Escherichia coli revealed that the genes for Shiga toxin and SLT-I were greater than 99% homologous; three nucleotide changes were detected in three separate codons of the A subunits. Only one of these codon differences resulted in a change in the amino acid sequence: a threonine in Shiga toxin at position 45 of the A subunit compared with a serine in the corresponding position in SLT-I. Furthermore, Shiga toxin and SLT-I had identical signal peptides for the A and B subunits, as well as identical ribosome-binding sites, a putative promoter, and iron-regulated operator sequences. These findings indicate that Shiga and SLT-I are essentially the same toxin. Southern hybridization studies with total cellular DNA from several Shigella strains and internal toxin probes for SLT-I and its antigenic variant SLT-II showed that a single fragment in S. dysenteriae type 1 hybridized strongly with the internal SLT-I probe. Fragments with weaker homology to the SLT-I probe were detected in S. flexneri type 2a but no other shigellae. No homology between the Shiga-like toxin II (SLT-II) probe and any of the Shigella DNAs was detected. Whereas SLT-I and SLT-II are phage encoded, no phage could be induced from S. dysenteriae type 1 or other Shigella spp. tested. These results suggest that the Shiga (SLT-I) toxin genes responsible for high toxin production are present in a single copy in S. dysenteriae type 1 but not in other shigellae. The findings further suggest that SLT-II genes are absent in shigellae, as are toxin-converting phages.

Cloning of genes for production of Escherichia coli Shiga-like toxin type II.

Genes controlling production of Shiga-like toxin type II (SLT-II) in Escherichia coli were cloned from the SLT-II-converting bacteriophage 933W and compared with the Shiga-like toxin type I (SLT-I) genes previously isolated and described from phage 933J. Subcloning analysis identified a region within the 4.9-kilobase EcoRI fragment of phage 933W that was associated with SLT-II production. Experiments with E. coli minicells containing these subclones demonstrated that the 4.9-kilobase EcoRI fragment encodes the structural genes for SLT-II. These experiments additionally showed the genetic organization of the SLT-II genes to be the same as that of the SLT-I genes, with the coding sequence for the large A subunit adjacent to that for the smaller B subunit. The mobilities of the SLT-II subunits in sodium dodecyl sulfate-polyacrylamide gels were slightly greater than those determined for the SLT-I subunits. Although apparent processing of the SLT-I subunits was observed with polymyxin B treatment of the labeled minicells, no processing of the SLT-II subunits was detected. Southern blot hybridization studies suggested that the DNA fragment carrying the SLT-II structural genes shares approximately 50 to 60% homology with the DNA of the SLT-I structural genes.

Two toxin-converting phages from Escherichia coli O157:H7 strain 933 encode antigenically distinct toxins with similar biologic activities.

Escherichia coli O157:H7 strain 933 contains two distinct toxin-converting phages (933J and 933W). The biologic activities and antigenic relationship between the toxins produced by 933J and 933W lysogens of E. coli K-12, as well as the homology of the genes that encode the two toxins, were examined in this study. The 933J and 933W toxins, like Shiga toxin produced by Shigella dysenteriae type 1, were cytotoxic for the same cell lines, caused paralysis and death in mice, and caused fluid accumulation in rabbit ileal segments. The cytotoxic activity of 933J toxin for HeLa cells was neutralized by anti-Shiga toxin, whereas the activity of 933W toxin was not neutralized by this antiserum. In contrast, an antiserum prepared against E. coli K-12(933W) neutralized 933W toxin but not 933J toxin or Shiga toxin. For E. coli 933, most of the cell-associated cytotoxin was neutralized by anti-Shiga toxin, whereas most of the extracellular cytotoxin was neutralized by anti-933W toxin. However, a mixture of these antisera indicated the presence of both toxins in cell lysates and culture supernatants. Among 50 elevated cytotoxin-producing strains of E. coli, we identified 11 strains isolated from cases of diarrhea, hemorrhagic colitis, or hemolytic uremic syndrome that produced cell-associated cytotoxins which were neutralized by the 933W antitoxin. Southern hybridization studies showed that the cloned toxin structural genes from phage 933J hybridized with DNA from phage 933W under conditions estimated to allow no more than 26% base-pair mismatch. These findings indicate that E. coli produces two genetically related but antigenically distinct cytotoxins with similar biologic activities which we propose to name Shiga-like toxins I and II. Strains of E. coli that produce elevated levels of Shiga-like toxin I or Shiga-like toxin II, or both, have been associated with the clinical syndromes of diarrhea, hemorrhagic colitis, and hemolytic uremic syndrome.

Characterization of monoclonal antibodies against Shiga-like toxin from Escherichia coli.

Three monoclonal antibodies, designated MAb 16E6, MAb 13C4, and MAb 19G8, were produced which recognize Shiga-like toxin (SLT) from Escherichia coli. All three monoclonal antibodies neutralized the cytotoxicity of E. coli SLT and were able to immunoprecipitate intact labeled toxin with Staphylococcus aureus protein A. The three antibodies were of the G1 heavy and kappa light chain classes. MAb 16E6 bound to the B subunit of SLT in Western blots and also neutralized the lethality of the toxin for mice and the enterotoxicity of the toxin in ligate rabbit ileal loops. The ability of MAb 16E6 to neutralize the cytotoxicity, lethality, and enterotoxicity of E. coli confirms the hypothesis that all three activities are associated with a single toxin. MAb 16E6 and MAb 13C4 also neutralized the cytotoxicity of purified Shiga toxin from Shigella dysenteriae type 1 and Shiga-like toxic activities in crude cell extracts from Shigella flexneri, Vibrio cholerae, Vibrio parahaemolyticus, and Salmonella typhimurium. Thus, Shiga toxin and the SLTs from E. coli, Shigella flexneri, V. cholerae, V. parahaemolyticus, and Salmonella typhimurium share a common B subunit epitope that is involved in neutralization. MAb 13C4 has been successfully used in a colony blot assay for the detection of bacterial colonies which produce high levels of SLT. Sixty-two different strains of bacteria were tested by both the cytotoxicity and colony blot assays for the presence of SLT. The colony blot assay detected all strains of bacteria which produce greater than or equal to 10(5) 50% cytotoxic doses of SLT per ml of sonic lysate. There were no false-positive results among the 62 samples tested.